We build a lot of screens that show numbers, and the rule that shapes all of them is the same: a number you cannot defend is worse than no number.
The newest one tracks the towns we have not opened yet. For each candidate town it shows how many verified tradespeople each trade needs before the town can open, how many are there now, and, where the evidence allows, roughly how long closing the gap might take at the recent rate of signups.
The interesting part is the phrase "where the evidence allows". If fewer than five tradespeople verified in the last ninety days, no estimate appears. Not a cautious estimate, not a wide range. Nothing, plus a sentence saying exactly why: below five observations, one outlier owns the answer, and quoting it would present noise as evidence.
When an estimate does appear, it states its own basis in the same breath: how many verifications, over how many days, and that it is a projection from a recent rate, not a promise. The document we hand to a letting agent asking "when will you cover us?" says the same thing the internal screen says, because they are generated from the same numbers.
If you work a trade in a town we have not opened, this is the machinery deciding where we knock next.